Plug-and-Play Sensor Evolution in Software-Defined Vehicles Through Asset Administration Shells

Abstract

Software-defined vehicles (SDVs) promise continuous software evolution, but sensor integration in deployed vehicles still depends on manual configuration, hardware-specific drivers, and repeated application-side adaptation. Existing standards address only parts of this problem. AUTOSAR standardizes selected sensor interfaces, the Vehicle Signal Specification (VSS) standardizes signal semantics, and Asset Administration Shells (AAS) standardize digital asset descriptions. However, these building blocks are rarely combined into an end-to-end onboarding path for sensors attached after vehicle deployment. This paper presents an AAS-driven plug-and-play sensor integration framework for safety-supervised sensors. Each sensor is represented by an AAS that combines hardware identification, interface metadata, VSS mappings, and references to containerized driver and optional application software. After physical connection, an edge gateway identifies the device, retrieves the matching AAS, deploys the referenced software, and publishes the resulting data under stable vehicle signal paths. A prototype on a resource-constrained edge gateway connected to an RR100 UGV demonstrates automated integration of a USB camera and an ultrasonic sensor. Across repeated integration cycles, the prototype reaches operational state in about 2 s, shows no failed events, and keeps applications independent of hardware-specific interfaces. The results indicate that the introduction of AAS as an executable onboarding contract enables automated runtime deployment of sensor drivers in SDV environments.